Thanks for forwarding this to Debian as well! I suppose until this gets reviewed and approved in Debian, I'm willing to sponsor this patch to the Ubuntu archives. The patch itself makes sense and the code is well written. I like it because, even though in normal operation such carefulness shouldn't be required, it's an additional safety net to make sure the ioctl will not block forever.
Before I proceed with uploading though, do you think you could think of some good test-case or reproduction steps for those not using g15daemon? I know this seems to be happening occasionally in that particular use-case, but it would be nice if a way of reproducing the original problem for reviewers was available. I imagine that might be a bit tricky to do.
Thanks for forwarding this to Debian as well! I suppose until this gets reviewed and approved in Debian, I'm willing to sponsor this patch to the Ubuntu archives. The patch itself makes sense and the code is well written. I like it because, even though in normal operation such carefulness shouldn't be required, it's an additional safety net to make sure the ioctl will not block forever.
Before I proceed with uploading though, do you think you could think of some good test-case or reproduction steps for those not using g15daemon? I know this seems to be happening occasionally in that particular use-case, but it would be nice if a way of reproducing the original problem for reviewers was available. I imagine that might be a bit tricky to do.